Method and Device for Imaging of Lensless Hyperspectral Image

1. A method of reconstructing a hyperspectral image, the method comprising: receiving an image photographed through only a single diffractive optical element attached to a single bare image sensor, wherein an imaging lens includes the single diffractive optical element; and reconstructing a hyperspectral image of the received image based on the received image and information about a point spread function for each wavelength of the diffractive optical element, wherein the single diffractive optical element has a diffractive rotation property that spectrally rotates a shape of the point spread function.

2. The method of claim 1, wherein the single diffractive optical element is configured to generate an anisotropic shape of the point spread function that varies with a spectrum.

3. The method of claim 1, wherein the reconstructing of the hyperspectral image includes: reconstructing the hyperspectral image of the received image based on the information about the point spread function for each wavelength, the received image, and a neural network trained by a previously generated learning model.

4. The method of claim 3, wherein the neural network is configured to learn a spatial prior and a spectral prior of the hyperspectral image by the learning model, and reconstruct a spectral image by diffractive rotation of the point spread function.

5. The method of claim 1, wherein the reconstructing of the hyperspectral image includes: reconstructing the hyperspectral image of the received image by reflecting an optimization technique for diffractive rotation of the point spread function.

6. The method of claim 5, wherein the reconstructing of the hyperspectral image includes: reconstructing the hyperspectral image of the received image by repeating the optimization technique a specified number of times.

7. The method of claim 3, wherein the neural network comprises multiphase neural network.

8. An apparatus for reconstructing a hyperspectral image, the apparatus comprising: a receiving unit configured to receive an image photographed through only a single diffractive optical element attached to a single bare image sensor, wherein an imaging lens includes the single diffractive optical element; and a reconstructing unit configured to reconstruct a hyperspectral image of the received image based on the received image and information about a point spread function for each wavelength of the single diffractive optical element, wherein the single diffractive optical element has a diffractive rotation property that spectrally rotates a shape of the point spread function.

9. The apparatus of claim 8, wherein the single diffractive optical element is configured to generate an anisotropic shape of the point spread function that varies with a spectrum.

10. The apparatus of claim 8, wherein the reconstructing unit is configured to reconstruct the hyperspectral image of the received image based on the information about the point spread function for each wavelength, the received image, and a neural network trained by a previously generated learning model.

11. The apparatus of claim 10, wherein the neural network is configured to learn a spatial prior and a spectral prior of the hyperspectral image by the learning model, and reconstruct a spectral image by diffractive rotation of the point spread function.

12. The apparatus of claim 8, wherein the reconstructing unit is configured to reconstruct the hyperspectral image of the received image by reflecting an optimization technique for diffractive rotation of the point spread function.

13. The apparatus of claim 12, wherein the reconstructing unit is configured to reconstruct the hyperspectral image of the received image by repeating the optimization technique a specified number of times.

14. The apparatus of claim 10, wherein the neural network comprises multiphase neural network.

15. A system for reconstructing a hyperspectral image, the system comprising: a single diffractive optical element attached to a single bare image sensor, wherein an imaging lens includes the single diffractive optical element; and a camera apparatus configured to take an image through only the single diffractive optical element, and reconstruct a hyperspectral image of the photographed image based on a point spread function information for each wavelength of the single diffractive optical element and the photographed image, wherein the single diffractive optical element has a diffractive rotation property that spectrally rotates a shape of the point spread function.

16. The system of claim 15, wherein the single diffractive optical element is configured to generate an anisotropic shape of the point spread function that varies with a spectrum.

17. The system of claim 15, wherein the camera apparatus is configured to reconstruct the hyperspectral image of the received photographed image based on the information about the point spread function for each wavelength, the received image, and a neural network trained by a previously generated learning model.

18. The system of claim 15, wherein the camera apparatus is configured to reconstruct the hyperspectral image of the received photographed image by reflecting an optimization technique for diffractive rotation of the point spread function.